International Journal of Legal Medicine

, Volume 127, Issue 2, pp 395–404

Toxicogenetics—cytochrome P450 microarray analysis in forensic cases focusing on morphine/codeine and diazepam

Authors

    • Institute of Legal Medicine, Forensic ToxicologyUniversity Medical Center Hamburg-Eppendorf
  • C. Augustin
    • Institute of Legal Medicine, Forensic Molecular BiologyUniversity Medical Center Hamburg-Eppendorf
  • T. Streichert
    • Department of Clinical ChemistryUniversity Medical Center Hamburg-Eppendorf
Original Article

DOI: 10.1007/s00414-012-0759-6

Cite this article as:
Andresen, H., Augustin, C. & Streichert, T. Int J Legal Med (2013) 127: 395. doi:10.1007/s00414-012-0759-6

Abstract

Genetic polymorphisms in cytochrome P 450 (CYP) enzymes could lead to a phenotype with altered enzyme activity. In pharmacotherapy, genotype-based dose recommendations achieved great importance for several drugs. In our pilot study, we ask if these genetic tests should be applied to forensic problems as a matter of routine. Starting from 2004 through 2008, we screened routine cases for samples where the relation of parent compound to metabolite(s) (P/M ratio), particularly morphine to codeine ratios and diazepam to its metabolites, was noticeable or not consistent with the information provided by the defendants. We found 11 samples with conspicuous results. These were analyzed for polymorphisms of the CYP 2D6 and 2C19 genes using the Roche AmpliChip Cytochrome P450 Genotyping test. If not previously conducted, a general unknown analysis by gas chromatography/mass spectrometry (GC/MS) was additionally carried out. For CYP 2D6, we found two cases with the genotype poor metabolizer (PM), three cases with heterozygote extensive metabolizer genotype classified as an intermediate metabolizer (IM) with probably reduced enzyme activities, but no ultrarapid metabolizer genotype. For CYP 2C19, two cases were characterized as IM phenotypes, with no PM found. Once we achieved no appropriate amounts of DNA, one case was excluded after GC/MS analysis. Only in one case could the polymorphism clearly explain the changes in drug metabolism. More frequently, a drug–drug interaction was thought to have a stronger impact. Additionally, our results suggest that IM genotypes may be more relevant than previously suspected. With respect to the small number of cases in which we thought a genotyping would be helpful, we conclude that the overall relevance of toxicogenetics in forensic problems is moderate. However, in some individual cases, a genotyping may provide new insight.

Keywords

PharmacogeneticsToxicogeneticsPolymorphismForensic toxicologyMicroarrayCYP 2D6CYP 2C19

Introduction

Most xenobiotics, including therapeutic drugs, are metabolized by the cytochrome P450 enzyme system. With respect to drug metabolism, the most important ones are CYP 2C9, 2C19, 2D6, 3A4, and 3A5. Clinically relevant polymorphisms of this enzyme system are described for CYP 2D6, CYP 2C19, and CYP 2C9 [1]. Typical drugs affected in their metabolism are, for example, codeine and tramadol for CYP 2D6, valproic acid for CYP 2C9, and diazepam and amitriptyline for CYP 2C19 [2].

Polymorphisms in these genes can change the enzyme activity, ranging from a complete deficiency to an ultrafast metabolism. These differences in drug metabolism could lead to severe toxicity or therapeutic failure by altering the relationship between the dose and the blood concentration of the pharmacologically active parent drug or metabolite, respectively. Although the enzymes could carry a number of different genetic polymorphisms, they can be categorized into four groups according to their associated phenotype: ultrarapid metabolizer (UM) containing multiple copies of the gene, extensive metabolizer (EM) with a single wild-type copy (the “normal” metabolizer), intermediate metabolizer (IM; heterozygotes) exhibiting decreased enzymatic activity, and poor metabolizer (PM; homozygotes) with no detectable activity [3].

For pharmacotherapy, the genotyping of cytochrome P450 genes is beginning to gain importance. By defining the genotype, the dosing of an administered drug can be individually adapted [47]. In the field of occupational health, the individual vulnerability of workers to certain chemicals can be predicted [810].

There are also some case studies with promise for forensic applications: The authors report fatal cases due to enzyme polymorphism, especially defective CYP 2D6 genotypes [1113], or ultrarapid CYP 2D6 metabolism [14, 15]. Though, in most of these cases, not only an altered metabolism but also drug–drug interactions played a relevant role.

In the field of postmortem toxicology, some early studies were carried out to determine the impact of polymorphisms in routine cases. Most of these studies focused on forensic cases which involved certain drugs, for example, amitriptyline [16], oxycodone [17], tramadol [18], or citalopram [19]. In one such study, cases with fatal drug intoxications were compared to two control groups of suicides and cases with natural causes of deaths, respectively [20]. In another study, genotyping was carried out in forensic cases which were suspicious because of their parent drug-to-metabolite ratio being unexpectedly high. In this study, only genotypes leading to CYP 2D6 poor metabolism were investigated [21].

These postmortem studies concluded that polymorphisms play no significant role in fatalities [16]; the risk of “PMs-pitfall” seems not to be very high [21], and drug–drug interactions were discovered to be more important than polymorphisms.

Despite these findings, Jannetto et al. [17] and Jin et al. [22] named pharmacogenomics as “molecular autopsy.” Musshoff et al. stated: “Although investigations and reports on genetically determined fatalities are still sparse up to date, modern biomarkers are expected to become a useful tool for forensic medicine in the future” [23]. Levo et al. predicted that genotyping applied to drugs of higher toxicity will aid in the interpretation of poisoning cases [18]. Sajantila et al. have the opinion that, at this stage, postmortem pharmacogenetics is, in general, still not sufficiently developed for courtroom purposes, although in some individual cases, it may provide new insight into the cause or manner of death of an individual [24].

In summary, there are some possible applications for genotyping in forensic issues. These questions include an uncommon parent drug/metabolite pattern or high concentrations of drugs despite low doses administered or prescribed, respectively. Genotyping might also help to estimate the time point of drug application more precisely or may prove of value in cases where medical responsibility for injury or loss may arise from adverse drug reaction [16]. Several authors highlight the relevance of genetic disposition for determining the cause and/or manner of deaths, i.e., whether it was suicidal or accidental [17]. In some cases of driving under the influence of drugs (DUID), a genotyping could be meaningful [25], for example, to clarify if a subject has taken prescribed drugs according to a physician's direction or had overdosed on the medication. Further forensic problems are conceivable, such as the question of a diminished or restricted criminal responsibility, drug abuse, or an existing drug addiction.

At present, there is no routine genotyping in our department of legal medicine, and the question arose whether it is necessary to pay more attention to genetic polymorphisms in the field of forensic toxicology. If there are uncommon patterns of parent drug and metabolites (P/M ratios), influencing factors could be polymorphisms in the cytochrome P450 enzyme system. Instead of massive genotyping, we chose a cost-effective strategy by preselecting cases with noticeable P/M ratios for genotyping. We began this pilot study to get an overall picture of the relevance of polymorphisms in the forensic context.

Methods

Toxicological analysis

Blood or serum samples were analyzed by immunological assays (CEDIA, Microgenics/Thermo Fisher Scientifics) for relevant (illegal) drugs. Results above cutoffs were confirmed by gas chromatography/mass spectrometry (GC/MS) analysis: Morphine and codeine were analyzed after solid-phase extraction (Bond Elut Certify, Varian) as pentafluoropropionamide derivatives. Morphine and codeine glucuronides were determined after incubation with ß-glucuronidase (Helix pomatia, Roche). Benzodiazepines were identified and quantified via gas chromatography with an electron capture detector after basic fluid/fluid extraction. Additional substances were screened in a “general unknown” screening after alkaline solid-phase extraction (C18, Varian) via GC/MS.

From 2004 until 2008, we collected cases with a parent drug–metabolite ratio (P/M ratio) which was noticeable or not concordant with statements of the offenders. We especially focused on morphine/codeine ratios and on diazepam and its metabolites because polymorphism in the cytochrome gene could change their metabolism and because these drugs are frequently found in our cases. For a stated codeine intake, the P/M ratio was assessed as not concordant if the level of morphine was higher or equal to the codeine level (see cases I–V). If only traces of the major metabolite were detected, a drug intake shortly before blood sampling or death was assumed (see cases VI–XI).

Samples were either blood samples of living persons (with sodium fluoride as preservative) (cases I to IX) or postmortem peripheral blood samples (vena femoralis) without additives (cases X and XI). Samples were stored up to 4 years at 4–8 °C (I–IX) or −20 °C (X and XI) until DNA preparation.

DNA preparation

DNA was extracted using the QiaAmp Minikit (Qiagen) according to the recommendations of the manufacturer. DNA quantification was performed via real-time polymerase chain reaction (PCR) using the Quantifiler kit on an ABI 7500 Real-Time System.

Genotyping

DNA samples were analyzed using the Roche AmpliChip Cytochrome P450 Genotyping test for use on the Affymetrix GeneChip Microarray Instrumentation System (service provider) following the manufacturer's instructions (Roche document no. 7/2006 04591445001-0336) [26]. The same protocol was applied to one postmortem sample. The polymorphisms analyzed are listed in Table 1.
Table 1

Tested alleles by the AmpliChip Microarray

 

Tested alleles

Total number

CYP 2D6

*1, *2ABD, *3, *4ABDJK, *6ABC, *7, *8, *9, *10AB, *11, *14AB, *15, *17, *19, *20, *25, *26, *29, *30, *31, *35, *36, *40, *41

24

CYP 2D6 duplications

*1xn, *2xn, *4xn, *10xn, *17xn, *35xn, *41xn

7

CYP 2C19

*1, *2, *3

3

Subjects with three or more active alleles, i.e., carriers of CYP 2D6 gene duplications, were classified as UMs. Subjects with two active alleles (*1, *2, *9, *10, and *17) were classified as EMs, and subjects with two inactive alleles (e.g., *3, *4, *5, *6, *7, *8) were classified as PMs [7, 27].

In our study, we followed the IM categorization according to Kirchheiner et al.; therefore, *1/*41 and *1/*10 were classified as EM and *1/*4 as IM phenotypes. On the basis of this classification, we categorized CYP 2C19 *1/*1 genotypes as EM and heterozygote genotypes with one active and one completely deficient allele (*1/*2) as IM with possibly reduced activity [7, 28].

Results

Cases

During 4 years and out of more than 13,000 investigated routine cases, we found 11 cases with drug–metabolite ratios which were noticeable or not concordant with statements of the offenders, thus suggesting a genotyping for CYP 2D6 and 2C19. The cases were DUID (n = 7), criminal offenses (n = 2), and two fatalities (see Table 2).
Table 2

Results of toxicological analysis and type of forensic question for all cases

Case

Toxicological result

Type of offense/question

I

0.023 mg/L morphine, 0.004 mg/L codeine

DUID, regulatory offenses

II

0.102 mg/L morphine, 0.015 mg/L codeine, 0.039 mg/L morphine-gluc., 0.021 mg/L codeine-gluc.

DUID, regulatory offenses

III

0.015 mg/L morphine, 0.017 mg/L codeine, 0.099 mg/L morphine-gluc., 0.044 mg/L codeine-gluc.

DUID, regulatory offenses

IV

0.100 mg/L morphine, 0.100 mg/L codeine

DUID, regulatory offenses

V

0.054 mg/L morphine, 0.029 mg/L codeine, 0.035 mg/L benzoylecgonine

DUID, regulatory offenses

VI

1.5 mg/L diazepam, 0.04 mg/L nordiazepam, 0.64 g/kg ethanol

DUID

VII

0.372 mg/L codeine

DUID

VIII

6.4 mg/L diazepam, nordiazepam n.d., 1.23 g/kg ethanol

Criminal offense

IX

1.1 mg/L diazepam, 0.02 mg/L nordiazepam, 1.36 g/kg ethanol

Criminal offense

X

1.08 mg/L codeine, sibutramine, moclobemide, acetaminophen, bromazepam, morphine n.d.

Fatal intoxication

XI

0.16 mg/L codeine, 0.70 mg/L codeine-gluc., 0.014 mg/L morphine, 0.18 mg/L morphine-gluc., oxazepam, temazepam, zopiclone, acetaminophen

Fatal intoxication

n.d. not detected, gluc glucuronides, DUID driving under the influence of drugs

Case I–V: driving under the influence of drugs—heroin or codeine?

Five cases of regulatory offenses according to the German Road Traffic Act (§24a StVG) were analyzed because of their ambiguous drug–metabolite ratios: In cases I to IV, the suspected persons declared the intake of codeine, but the morphine/codeine relation pointed to a consumption of street heroin (Fig. 1). In case IV, the general unknown screening revealed meconine, which points to the incorporation of heroin. Thus, a genotyping was not reasonable because the stated codeine uptake could be disproved at this step.
https://static-content.springer.com/image/art%3A10.1007%2Fs00414-012-0759-6/MediaObjects/414_2012_759_Fig1_HTML.gif
Fig. 1

Metabolism of codeine

In case V, the accused declared to have taken heroin for the first time after a long period of abstinence. The ratio of morphine to codeine was unusual for a case of heroin consumption.

Case VI–VII: driving under the influence of drugs—overdose or therapeutic?

Case VI was a person who drove his car under the influence of alcohol (below the legal cutoff) and a high concentration of diazepam (1.5 mg/L). The metabolite nordiazepam could be detected only at a lower level (0.04 mg/L) (Fig. 2). For judging this case, it was important to know if the subject overdosed on the drug or had taken the prescribed drug according to a physician's direction.
https://static-content.springer.com/image/art%3A10.1007%2Fs00414-012-0759-6/MediaObjects/414_2012_759_Fig2_HTML.gif
Fig. 2

Metabolism of diazepam

In case VII, the person showed an altered behavior fitting to drug effects during a police check. He was stopped at 1:35 a.m. and had blood drawn about 1 h later. The suspected person conceded the intake of codeine several hours before; the results of the blood analysis indicated 0.372 mg/L codeine and no morphine, which pointed to an intake of this drug a short time before.

Case VIII and IX: criminal offense—drug abuse or medical sedation?

Case VIII was a criminal offense where the accused was obstructing a police officer. The blood analysis revealed 1.23 g/kg alcohol, 6.4 mg/L diazepam (no nordiazepam could be detected—limit of detection, LOD, was 0.011 mg/L). It had to be clarified if he abused the drug or if it was an administration by medical staff.

In case IX, the person attempted a car theft. In the blood, diazepam (1.1 mg/L), traces of nordiazepam (0.02 mg/L), and ethanol (1.36 g/kg) could be detected. No medical intervention after the event was documented. Due to the high ratio of diazepam to nordiazepam, an acute abuse was suspected, but a normal uptake of diazepam without abuse tendencies could not be excluded if he would be a PM.

Cases X and XI: fatal intoxications—time point of codeine use?

In case X, we found in the femoral vein blood 1.08 mg/L codeine but no morphine (LOD = 0.002 mg/L). Case XI displayed in the femoral vein blood 1.16 mg/L of codeine, 0.70 mg/L of codeine-glucurononides, 0.014 mg/L morphine, and 0.18 mg/L of morphine-glucurononides. Besides that, we found in both cases additional drugs (see Table 2). We asked if these ratios with high amounts of codeine could be explained by (1) a polymorphism leading to a poor metabolizer phenotype, (2) drug interaction, or (3) the time point of intake (shortly before death).

Genotyping

The results of the microarray-based genotyping are shown in Table 3. In four cases, we hypothesized an ultrarapid CYP 2D6 phenotype; in four cases, we supposed a poor CYP 2D6 phenotype, and three times, we expected to find a poor CYP 2C19 phenotype. Despite two attempts, in case XI, the material was not appropriate to isolate sufficient amounts of DNA for genotyping; in case VI, only the genotyping for CYP 2C19 was successful.
Table 3

Results of the microarray-based genotyping for CYP 2D6 and 2C19, our hypothesis according to the toxicological results, and the confirmation are displayed for all cases

Case number

Genotype 2D6

AmpliChip-predicted phenotype CYP 2D6

Genotype 2C19

AmpliChip-predicted phenotype CYP 2C19

Hypothesis

Confirmation

I

*1/*41

Extensive

*1/*1

Extensive

Ultrarapid 2D6

No

II

*4/*4

Poor

*1/*1

Extensive

Ultrarapid 2D6

Contrary

III

*1/*10

Extensive

*1/*1

Extensive

Ultrarapid 2D6

No

IV

 

Not tested

 

Not tested

  

V

*1/*4

Extensive (IM)b

*1/*1

Extensive

Poor 2D6

(Yes)

VI

 

No call

*1/*2

Extensive (IM)b

Poor 2C19

(Yes)

VII

*4/*4

Poor

*1/*1

Extensive

Poor 2D6

Yes

VIII

*1/*4a

Extensive (IM)b

*1/*1

Extensive

Poor 2C19

No

IX

*1/*4

Extensive (IM)b

*1/*1

Extensive

Poor 2C19

No

X

*1/*41

Extensive

*1/*2

Extensive (IM)b

Poor 2D6

No

XI

 

No DNA

 

No DNA

Poor 2D6

?

no DNA not enough DNA for genotyping, no call quality of signals was not sufficient to identify the genotype and predict a phenotype, not tested no genotyping was carried out

aAdditional CYP 2D6 genotyping by pyrosequencing

bIM phenotype classified according to Kirchheiner et al. [7]

For CYP 2D6, in four of nine cases, the genotyping indicated the extensive genotype (*1/*1, *1/*41, *1/*10) with normal enzyme activity. In two cases, we found a poor metabolizer genotype (*4/*4), and in three cases, we detected heterozygote genotypes (*1/*4) with a possibly reduced enzyme activity, which we classified as IM. For CYP 2C19, we found two cases with heterozygote genotype (*1/*2), classified as IM, and no PM was detected.

Discussion

Results of genotyping

In general, genotyping is a promising way to get more information in complex cases involving polymorphous enzymes and substances with unusual P/M ratios. In our study, we selected cases out of routine forensic cases in which a genotyping could be useful. During 4 years, we found only 11 appropriate cases which met the inclusion criteria. An exact comparison to the ratio of the studies mentioned earlier was not possible because in these studies, the majority of investigators were following other approaches. They selectively collected all cases which were related to one drug (e.g., amitriptyline [16], oxycodone [17], tramadol [18]) or with fatal drug intoxications [20], initially without regard to the P/M ratio. Therefore, the number of samples was higher in these studies (n = 202, 15, 33, and 242, respectively). Only Druid et al. had a similar approach compared to our pilot study. They also explored cases in which the P/M ratio was unexpectedly high (n = 22; control group, n = 24); however, they investigated PM genotypes only [21].

In our study, patients were separated into carriers of none (PM), one (IM), two (EM), or, in the case of CYP 2D6 gene duplication, into carriers of more than two (UM) functional alleles. In 1 of the 11 samples, no genotyping was carried out because it was unnecessary after further information, and in one case, no appropriate DNA extraction was possible. In two of these remaining nine cases with successful genotyping for CYP 2D6, we found a poor metabolizer genotype (22 %). In three cases, an intermediate metabolizer genotype for CYP 2D6 (33 %) was found, and in two cases of CYP 2C19, IM (22 %) was discovered. No ultrarapid metabolizer was found in our selection.

Only in three of these cases could the result of genotyping support our hypothesis (in case VII, we expected and found CYP 2D6 PM; in cases V and VI, we expected PMs, but found at least potentially reduced metabolism due to predicted IM phenotypes), according to 27 % of all 11 selected cases. In one case (II), the opposite genotype was found, and in three cases (VIII, IX, and X), the predicted IM phenotype had no relevance in the particular question.

The determination of an IM phenotype is not standardized so far. According to Kirchheiner et al. [7], we found three IMs (*1/*4). Following the definition of Zanger et al. [29], the genotypes *1/*41 and *1/*10 should be predicted as IMs too. Then, we would have three additional cases of CYP 2D6 IMs in our sample. Zhou et al. [30] determined only genotypes with one null and one defective allele as IM. Applying this definition to our results, we would have had no IM phenotypes.

In Caucasians, the incidence of CYP 2D6 polymorphisms is 5–10 % for PMs, 10–15 % for IMs, and about 1–10 % for UMs [3133]; the incidence of CYP 2C19 polymorphisms is 2–4 % for PMs [1]. Druid et al. found in 1 of 46 postmortem cases (study group, n = 22; control group, n = 24) a CYP 2D6 poor metabolizer genotype (2.2 %), but in 10 of 22 cases, they supposed IM genotypes (45 %); no UMs were tested in their study by PCR [21]. Levo et al. found 12 % CYP 2D6 UMs (4 of 33), 12 % PMs, and 27 % IMs [18]. Jannetto et al. [17] found 13 % CYP 2D6 PMs and 27 % IMs in their study, but analyzed no UMs [17]. In the study conducted by Carlsson et al., the percentage of poor metabolizer was 6 % for CYP 2D6 and 8 % for CYP 2C19 (n = 50) [34]. The same group found in an earlier study no PMs regarding CYP 2C19 and only two (3.8 %) PMs regarding CYP 2D6 (n = 53) [19] (see Table 4). Zackrisson et al. compared three causes of death (fatal intoxications, FI, n = 242), suicides without intoxications (S, n = 262), and natural deaths (ND, n = 212)). They found a higher number of UMs in the group of suicides, 4.7 % (vs 2.5 % (FI) and 0.5 % (ND)), but they had no explanation for these findings. They detected PMs in only 11 of the fatal intoxication cases (4.7 %). In the control group of ND, they found 8.5 % poor metabolizer genotypes [28]. Ahlner et al. found a tenfold higher frequency of individuals carrying more than two active CYP 2D6 alleles in violent suicide cases compared with the natural death cases [35].
Table 4

Genotyping results of postmortem studies

Study

Total number of cases

CYP 2D6

CYP 2C19

UM

IM

PM

IM

PM

Our study

9

0

3 (33 %)

2 (22 %)

2 (22 %)

0

Druid et al. [21]

22/46

n.t.

10 of 22 (45 %)

1 of 46 (2 %)

n.t.

n.t.

Levo et al. [18]

33

4 (12 %)

9 (27 %)

4 (12 %)

n.t.

n.t.

Jannetto et al. [17]

15

n.t.

4 (27 %)

2 (13 %)

n.t.

n.t.

Holmgren et al. [19]

53

n.t.

n.t.

2 (3.8 %)

n.t.

0

Koski et al. [16]

202

14 (7 %)

60 (30 %)

13 (6 %)

36 (18 %)

11 (5 %)

Carlsson et al. [34]

50

2 (4 %)

16 (32 %)

3 (6 %)

13 (26 %)

4 (8 %)

Prevalence in Caucasians

 

5–10 %a

10–15 %b

5–10 %c

No data

2–4 %a

The number (and percentage) of cases with UM, IM, or PM genotypes for CYP 2D6 and CYP 2C19 are displayed. For comparison, the prevalence of these genotypes in a Caucasian population according to literature is mentioned in the last line

n.t. not tested

aAccording to Rogers et al. [1]

bAccording to Bock et al. [47]

cAccording to Zanger et al. [29]

Calculation of significance statistics and comparing these results directly with results of other investigators are not admissible, due to the small number of cases investigated in our study. However, we calculated the percentile rates to estimate and compare the rate of different genotypes in our sample (see Table 4). The rate of CYP 2D6 PMs found in our study (22 %) was higher than in other studies and above the prevalence in a Caucasian population. The rate of predicted CYP 2D6 IM phenotypes is in line with other studies and is also above the prevalence. In contrast to other studies and to our expectation to find three UMs, we found not one. One reason could be the small sample which was investigated in our pilot study.

For CYP 2C19, no data for IM prevalence could be found, but the proportion of 22 % we found is comparable with the results of Koski et al. and Carlsson et al. (see Table 4) [16, 34]. As in the study by Holmgren et al. [19], we found no CYP 2C19 PM genotype, despite the fact that we expected this in 30 % of the tested cases. In the other two studies which also determined PMs for CYP 2C19, higher percentages of PMs were found, considerably above the prevalence in Caucasians.

Discussion of the cases

In our study, we chose the approach of selecting cases by their peculiar ratios of drug to metabolites or implausible statements regarding the drug intake. The question was whether the toxicological results were caused by an enzyme polymorphism or should be explained by other means.

Cases I to III

Due to their ambiguous ratios of morphine to codeine, we hypothesized in cases I to III an ultrarapid phenotype. This was not confirmed by the microarray analysis. Quite the contrary was found in case II: Indicated by the poor metabolizer genotype, it is most likely that suspect II consumed not only codeine but both substances, morphine and codeine, or even heroin. Heroin intake is possible for case I and III as well. However, an ingestion of codeine could also not be excluded: Supposedly, these persons were in the final phase of codeine elimination. There are hints that during this phase, the ratio of morphine to codeine converges or turns into the opposite. This was not described for blood [3638] but for urine samples only [39]. In one study, 24 h after uptake of 30 mg codeine, the ratio of morphine to codeine was above 1 in UM but not in EM. However, these results were related to the ratio of total morphine to total codeine. Regarding the free compounds, the morphine to codeine ratio was below 1 during 24 h [40]. Another reason for the constellations in cases I and III could be a UGT 2B7 polymorphism [41], but the ratio of free-to-bound morphine or codeine gives no clue to a decreased UGT activity in case III. In case I, no determination of total opiates had been carried out. Of course, explanations other than polymorphisms are possible: The metabolism could be affected by enzyme inhibition or induction caused by other drugs, illegal substances, herbal drugs, nutrition, alcohol, or tobacco. One has to take into consideration that in case I and III, an inhibition of the CYP 3A4 degradation path (codeine to norcodeine) could explain the P/M ratios. Genetic variations for CYP 3A4 isoforms are described, though it is unclear if they lead to changes in the phenotype and the metabolism. Inhibitory effects on CYP 3A4 may also be caused by macrolides (erythromycin), immunosuppressant drugs like cyclosporine, antidepressants like selective serotonin reuptake inhibitors (e.g., paroxetine), or drugs for HIV therapy [2, 42]. We did not find any suspect substances in the general unknown screening, but of course, this screening would not detect all of the above-mentioned agents.

Case V

In case V, we searched for an explanation for a slower degradation of codeine after heroin consumption because of an unusual P/M ratio. The genotyping could not support our hypothesis of the poor 2D6 metabolizer, but we found a homozygote genotype, supposing an intermediate metabolizer type with a reduced metabolic ratio. In addition, in this case, the codeine metabolism could possibly be decreased because of the intake of an antibiotic. Clarithromycin and erythromycin, for example, are effective inhibitors of CYP 3A4, which is also involved in codeine metabolism (see Fig. 1).

The use was indicated in the morning, and the blood sample was taken at 6 p.m. The suspect stated also an intake of acetylsalicylic acid in the morning and of a broad-spectrum antibiotic twice a day (neither were investigated by the toxicological analysis).

Case VI

In case VI, we suspected a diazepam overdose shortly before blood sampling because we detected only traces of the metabolite nordiazepam. Due to the genotyping, no poor metabolizer genotype, but a CYP 2C19 intermediate phenotype, was found, possibly responsible for a reduced metabolism. Therefore, an accumulation of diazepam, despite a correct ingestion of this drug, could not be excluded.

Case VII

In case VII, our hypothesis of a CYP 2D6 poor metabolizer genotype could be confirmed, discharging the person of the suspicion of an intake of codeine a short time before driving.

Case VIII

In case VIII, a criminal offense, we detected high concentrations of diazepam but no metabolites, so we suspected a 2C19 poor metabolizer. The microarray analysis revealed an EM genotype for 2C19. However, this could not explain the absence of nordiazepam, one of the metabolites of diazepam. However, it is striking that other metabolites like temazepam or oxazepam, which are metabolized by CYP 3A4 (Fig. 2), were not detectable either. Again, we could not exclude an inhibition of CYP 2C19 and CYP 3A4 by substances we were not searching for. The supposed IM phenotype for CYP 2D6 had no relevance for the present question. The very high concentration of diazepam could also be a hint for pre-analytical problems like incorrect blood drawing (e.g., administration of drug and blood drawing over the same venous line). Most likely, the offender was given the drug by medical staff as a sedative during the period of fixation, shortly before the blood sample was drawn.

Case IX

In case IX, higher concentrations of diazepam (1.1 mg/L), but no relevant amount of its metabolites, were detected in the blood sample, leading to the hypothesis of a reduced metabolism due to a CYP 2C19 poor metabolizer genotype. Genotyping revealed that case IX was a 2C19 extensive metabolizer. The supposed IM phenotype for CYP 2D6 had no relevance for the present question. In his questioning, the suspect declared having taken methadone, buprenorphine, and doxepine. Regarding drug–drug interactions, buprenorphine could act as a strong reversible CYP 3A4 inhibitor [43], possibly interfering with the degradation of diazepam to temazepam (see Fig. 2), and methadone could competitively inhibit the way via CYP 2C19 (diazepam to nordiazepam). However, neither buprenorphine nor methadone was found in full toxicological analysis.

Case X

In case X, we assumed a CYP 2D6 poor metabolizer because of the high concentrations of codeine in blood and brain tissue without detectable concentrations of morphine. This could be explained by an intake shortly before death or by an altered metabolism. The genotyping pointed to a CYP 2D6 extensive metabolizer; therefore, a reduced metabolism ratio by genetic disposition could be excluded. Apart from that, we found sibutramine, bromazepam, and acetaminophen, all substances that are metabolized by CYP 3A4, possibly affecting the conversion from codeine to norcodeine and moclobemide, known as an inhibitor of CYP 2D6. These results could possibly explain the high concentration of codeine in these samples. The supposed IM phenotype for CYP 2C19 had no relevance for the present question.

Discussion of the method

In 1 of the 11 samples, we had problems with sufficient DNA preparation for analysis. In another case, the quality of signals was not sufficient to identify the genotype and to predict a phenotype. This is comparable to, but slightly higher than, other studies [16, 21, 44]. Considering that not all authors reported the rate of unsuccessful DNA preparation and no sufficient data for microarray techniques are available, it seems that PCR methods might be slightly more robust for genotyping than the microarray technique.

The main advantage of using microarrays for genotyping is the simultaneous analysis of 24 alleles in the CYP 2D6 gene, as well as seven duplications and three alleles in the CYP 2C19 gene (see Table 1). Of course, microarray analysis is limited to mutations mapped on the microarray. Other mutations with possible impact on the phenotype could be overlooked.

Relevance of polymorphism in forensic context

Regarding our results, the ratio of polymorphisms with forensic relevance in our routine cases was not very high. Certainly, in our study, we detected only a fractional amount of all CYP 2D6 or 2C19 polymorphisms in our patient population. One reason for this is that not in all forensic cases with persons who have a polymorphous enzyme are drugs involved which are metabolized by this enzyme. Another reason is that a lot of drugs are degraded by more than one CYP isoform; therefore, the alternative method of metabolism could compensate for the reduced activity. Furthermore, according to our study design, we preselected cases in which we expected a polymorphism, and in addition, we thought that this genetic disposition would have influenced the P/M ratio of the detected drugs morphine/codeine or diazepam. Aside from this, other relevant drugs with polymorphous enzymes involved in their metabolism, mentioned for example by Sajantila et al. (e.g., tramadol, amitriptyline, doxepin, fluoxetine, imipramine, and venlafaxine [24]), were not included in our preliminary study.

Therefore, we certainly did not select all individuals potentially carrying a polymorphism. But, due to our hypothesis, we wanted to check if genotyping could help to interpret cases with inconsistencies between toxicological findings and anamnesis.

In our study, we expected a poor metabolizer genotype in eight cases, but only in one of these could our hypothesis be confirmed. In the case of the second PM, we had expected the opposite genotype (UM). As one result, it was noticeable that the part of heterozygous genotypes with possibly reduced metabolic capacity was rather high in all studies (with percentages from 27 to 45 % for CYP 2D6 and from 18 to 26 % for CYP 2C19; see Table 4). We found IM phenotypes in five cases; in two cases, we supposed a relevant effect of this polymorphism. Several pharmacokinetic studies suggest that the drug oxidation capacity of the IM phenotype is severely reduced and may be comparable to that of PMs, especially under conditions of long-term treatment [29]. Therefore, it might be possible that this kind of genotype is of higher importance than believed thus far.

Drug–drug interactions

In the study by Druid et al., a number of possible drug interactions were found which could explain some of the unexpected high metabolic ratios for certain substances [21]. Zackrisson et al. concluded that in general, drug interactions seem to be of greater importance than genetically determined deficiencies in the metabolic capacity [20]. Even in our cases, the relevance of possible drug–drug interactions was obvious. Unfortunately, not in all of these cases could drug–drug interactions be investigated properly. Therefore, in principle, an entire toxicological analysis including a general unknown screening should be carried out in cases of forensic relevance to detect drugs which could be relevant for interactions or point to the origin of drugs found in the specimen (e.g., in case IV: meconine from street heroin). Certainly, it should be kept in mind that not all relevant substances, for example, the ingredients of a grapefruit or some antibiotics, will be detected by routine methods. Other influences, like heavy tobacco consumption (leading, for example, to pharmacokinetic interactions with drugs that are CYP 1A2 substrates, such as clozapine, fluvoxamine, olanzapine, and theophylline [45]), could be excluded by further tests. Additionally, other conditions affecting the pharmacokinetics, and thus leading to accumulation of drugs or their metabolites, must be considered (e.g., age, severe diseases, renal and liver function) [46]. Moreover, additional factors affecting an individual's pathophysiological phenotype related to drug efficacy were mentioned, for example by Sajantila et al. (e.g., gender, hormonal changes, environmental toxins) [24].

Conclusion

Contrary to a relevant prevalence of polymorphisms in Caucasians, we found only a small number of conspicuous results during the period of examination (11 cases). In ten of these cases, a genotyping should be carried out, but only in nine of these cases was the quality of the DNA sufficient for a genotyping by microarray analysis. The hypothesized genotype could be confirmed in one of these cases, and for one case, we found a contrary result. The rate of intermediate metabolizers was unexpectedly high in our sample, indicating that this kind of polymorphism is more important than supposed until now.

Although the power of our study was limited, we conclude that the overall relevance of toxicogenetics in forensic problems is moderate. At present, we would not recommend routine genotyping in forensic cases, even though, in selected questions, it would be useful for the validation of analytical findings. However, one should bear in mind that many genes are still unstudied for the forensic context.

It seems that drug–drug interactions are more appropriate to explain implausible ratios of parent compound to metabolite, supporting the request for full toxicological examination in forensic cases and the holistic interpretation of toxicological results with respect to the case history.

Acknowledgments

We thank Kristin Klätschke for handling the DNA samples.

Copyright information

© Springer-Verlag 2012